497

Smart Artifacts Mediating Social Viscosity

Juan Salamanca

Every afternoon hundreds of pedestrians scramble to cross the five intersecting streets of the Shibuya district in Tokyo each time the traffic lights grant them the right of way over motor vehicles. Astonishing collective choreographies of coop-erative and collaborative interactions emerge from walkers, each with their own destination, negotiating trajectories to make it safely across the street. In doing so they encounter not only each other but also several nonhuman actors such as crosswalks, sidewalks, and signs, constituting a network of artifacts and crowds.

This kind of naturally occurring interaction in the wild exhibits recurrent forms of emergent and dissolving interactions that may inform effective method-ologies for the design of “socially apt” computational artifacts. A careful analysis of how artifacts participate in micro- social interactions serves to elucidate a novel approach for the design of artifacts that actively mediate people’s unplanned coop-eration, collaboration, conflict, or negotiation. The method used in this research picks emerging interactions mediated by regular artifacts, then replaces the arti-facts with a smart version of them programed to foster the same interactions, and finally contrasts the observations to derive conclusions about how to design such smart artifacts. To do so, this research proposes a triadic unit of analysis that em-braces human and nonhuman actors for the examination of actors’ flow of action in such mediated interactions. The triadic unit of analysis serves as a methodolog-ical instrument that unpacks action’s meaning and intentionality assembled when humans and artifacts intertwine their programs- of- action (Latour 1999). Meaning and intentionality can be inferred from the expected outcomes of parties involved in arranged interactions. For example, one person collaborating with someone else to move a bulky piece of furniture can infer the meaning and intention of her partner’s effort because their actions aim for the same result. But this research is interested not in such prearranged interactions but in emergent ones where par-ties are unaware of what others are aiming to, thus the meaning and intentionality of their actions can be inferred only from moral values while their actions unfold. In the formulation of this methodological instrument, I dislocate artifacts from being regarded as the equipment for human action and recast them as social ac-tors with scripted programs- of- action that participate in social collectives (Callon and Law 1995; Latour [1991] 1993, 2005). This conceptual move allows the adoption of the triadic unit of analysis, in contrast to the traditional user- object dyad that figures centrally in the User- Centered Design paradigm. From early empirical stud-ies using this analytical instrument I observed that cooperative and collaborative

498 SALAMANCA

interactions have some sort of internal resistance to collective action flow that re-sembles how a fluid with certain viscosity flows. According to my analysis I define social viscosity as the resistance to action flow exerted by actors— whether human or nonhuman— acting concurrently. Such resistance is created by mutual distur-bances that occur between humans and artifacts during the interactions that bind them together in a social practice.

The case presented herein compares the social dynamics of walkers in a regu-lar crosswalk with those of walkers in a custom- designed smart crosswalk. It illustrates how interactive aspects of the smart crosswalk’s design promote mu-tually beneficial interactions between pedestrians crossing two intersecting streets. The proposed smart crosswalk is not intended to be a future infrastruc-ture available on streets. It is rather a proof of concept that helps me to explore the benefits and drawbacks of intervening unplanned collective action flows with arti-facts that intend to help people do better together. From a first analysis of pedestri-ans in the wild, I identified the traditional crosswalk (the graphical pattern of parallel white bars painted on the ground) as a relevant actor that constitutes, to-gether with the traffic lights and sidewalks, a stage for unplanned interactions be-tween strangers and cliques of acquaintances. Next, I designed a smart crosswalk programmed to interpret the configurations and behaviors of cliques of walkers walking in opposite direction, and adaptively promote cooperation and collabora-tion. Then, I re- created observed walkers’ interactions in a laboratory setting with selected participants under several smart artifact conditions and compared their action flows. From the comparison, I derive that the social viscosity of a network could be partially determined by ways in which the design of smart artifacts signi-fies the dominant moral agreements of the practice in which they partake. I con-clude with a set of methodological considerations for the design of socially apt artifacts oriented to promote balanced social interaction by means of computa-tional artifacts.

The Design of Smart Artifacts

A necessary starting point for my analysis is to determine to what extent an arti-fact can be defined as smart. I extend Maturana’s (2002) notion of structured de-termined systems to define smart artifact as a tangible computational unit that reacts to external phenomena depending on the structure and organization of its constituent elements. From a posthumanistic position, artifacts can be perceived and interpreted by people as social actors that participate in— hinder or potentiate— their courses of action. In the same vein, smart artifacts incorporate technologies for “perceiving” and “interpreting” people’s actions and contextual information, in order to adapt themselves to the social context (Salamanca 2012). In its most simple composition, the structure of a smart artifact has three core elements: (1) tangible components— hardware, electrical parts, sensors, and actuators; (2) soft-ware components— operational system and programs; and (3) connectivity components— ports, antennas, and communication protocols (Porter and Heppel-mann 2014). I am particularly interested in smart artifacts that foster collective action in social settings such as intelligent vehicles, adaptive traffic signals, public screens, sensible floors, and crowdsourced products and services (e.g., collabora-tive digital cartography such as Waze traffic maps), to mention a few instances of social computing technologies.

499SMART ARTIFACTS

Computational artifacts have been the subject of inquiry of the interaction de-sign community for quite a while. In 2011, the Museum of Modern Art in New York exhibit Talk to Me: Design and the Communication between People and Objects high-lighted nearly 200 design projects that explore communication between people and computational artifacts (Antonelli 2011). According to Antonelli, the collection portrayed how contemporary designers not only shape form, function, and mean-ing, but also devise algorithms that enable pancommunication between every-thing and everybody in all possible combinations (Rosenberg 2011). The exhibition brought forth strong evidence of how the object of inquiry in interaction design evolved from how to make computational machines actionable and understand-able, to how computational artifacts embody personalities that fit unobtrusively into people’s everyday practices. The Talk to Me exhibit urged interaction designers to look beyond the understanding of users’ cognition, psychology, language, and action, central aspects of the User- Centered Design approach. This design para-digm, so popular among designers during the last three decades, proved useful to convert enigmatic computing machines into consumer products (Norman and Draper 1986; Norman 2002), but it now falls short in accounting for how pervasive computational artifacts actually mediate people’s social practices. This research seeks to contribute to a better understanding of the ways that smart artifacts worsen or ameliorate social viscosity in order to create socially apt artifacts.

Hybrid Participation in Social Practices

While the User- Centered Design perspective denies the ascription of action inten-tion to artifacts by assuming that humans— the “doers”— compose the totality of the system’s agency, a practice- centered design is open to distributing the system’s action— the “doing”— in collectives of actors that include artifacts. The latter not only extends the scope of the human information- processing model in interaction design, but also anchors a strand of research concerning the design of things that participate in society as assistants, agents, delegates, signifiers, mediators, or in-termediaries. Forlano (2016) argues that in decentering the human and embracing the attribution of agency to nonhumans, some critical and speculative design practices (Dunne 2005; Dunne and Raby 2013; Ratto 2011) are challenging the boundaries of the corporate design brief by eliciting points of tension among de-sign traditions. In the same vein, my proposal of recentering design practice opens into fundamental contributions that are primarily oriented toward expanding the domain of interaction design from user interface and experience to adaptive me-diators in social practices. On similar grounds Kuutti and Bannon (2014) propose a turn to practice in human- computer interaction. For them, a practice perspective encompasses a broader domain beyond the human- computer dyad, and brings to center- stage bodies, environment, artifacts, human capabilities, motivations, and values. Moreover, Knorr Cetina (1997) concurs with some posthumanist thinkers (Callon and Law 1995; Latour 1988) in that the analysis of a social practice is im-perfect unless it accounts for how artifacts mediate human activities. They argue that we are increasingly living in an object- centered society where artifacts take on significance through their roles not only as equipment but also as mediators, elements of cohesion, or commonplaces of identity characterization. Following this strand of thought, artifacts are not merely tools or commodities, but active partakers in human- artifact collectives.

500 SALAMANCA

Material Mediation of Social Practices

Material mediation in social practices occurs when artifacts participating in inter-actions “displace the actors’ viewpoints, act as a memory support and contribute to the setting up of a work collective” (Vinck 2012, 97). In the argument presented herein, mediating smart artifacts are envisioned as “social signifiers” (Norman 2011) of the dominant form of sociality of the practice mediating the dissonances created when individuals engage in unplanned social interactions (Serrano [1977] 2008).

Even though smart artifacts might not be entitled to privileges and obligations in the same manner as humans, they can be designed to reify social patterns, motives, moral principles, and political agendas (DiSalvo 2012). Latour (1999) famously uses a speed bump as an example of how things reify such aspects of human sociality. When a speed bump located on a neighborhood street compels drivers to slow down, it reifies the desire of the neighbors to keep their children safe from cars and represents the neighbors’ authority over streets, even though drivers are entitled to drive freely in the city. It is in the scope of activities interre-lating several actors— parents caring for their children, children playing on the street, and drivers commuting in their cars— that the speed bump reveals its moral facet because in this moment the bump behaves as an actor mediating the actions of multiple parties. Such scope of “embodied, materially mediated arrays of human activity centrally organized around shared practical understanding” is what (Schatzki 2001, 2) defines as a social practice.

Suchman’s canoeing example (Suchman 2007, 72; Hutchins 1995) helps us to illustrate the point. Situated action contends that artifacts and human actions are meaningful in relation to their moment- by- moment circumstances rather than to plans per se. A canoeist dealing with the challenges posed by rapids resorts to whatever embodied skills he or she has rather than carefully monitoring the sta-tus of a detailed plan. During the course of action the canoeist turns to the devised plan to identify the most convenient position to be able to face the challenges an-ticipated before boarding. In naturally occurring practices, such as walking in public spaces, people often throw themselves into acting purposefully without de-tailed courses of action. But, if we look at the distributed agency of the sociotech-nical configuration of the street, a clique of walkers is as subscribed to the program- of- action of a sidewalk as the canoeist is subscribed to the current. Thus, both sidewalk (a designed artifact) and river (a natural object) codetermine the ac-tion of any other actor entangled in their respective programs. The action flow of each pedestrian in a clique can be estimated (by herself, an observer, or a smart artifact) in relation to the sidewalk’s design. But the fluidity and consistency of their action flow depends on those of the other members of the clique; those of other humans and nonhumans acting concurrently on the sidewalk.

Unpacking Assembled Meaning and Intentionality: A Methodological Instrument

Up to this point I have argued that the analysis of artifacts participation in social practices is a source of insights for the design of socially apt smart artifacts that mediate emergent interactions. Such analysis looks at how the programs- of- action

501SMART ARTIFACTS

of artifacts codetermine those of humans resulting in a network of entangled ac-tors. In this section I introduce a methodological instrument to investigate the as-sembling of meaning and intentionality of action in those entanglements.

As actor- network theory suggests, the methodological approach in this research is to follow the action of networked actors, whether human or nonhuman (see, e.g., (Latour 1999, 2005; Callon and Law 1995). To do so, I formulate a triadic unit of anal-ysis constituted by a human actor associated with an artifact, both interacting as a whole with another human actor or collective (see figure 1).

The minimal requirement to be regarded as an actor of the triad is to get sub-scribed to the ongoing course of action. This means that new actors are participants in the network when they influence the course of action of current actors. While actors act on their own, the collective outcome is not necessarily a competitive game where each one maximizes one’s payoffs, but a stable imbroglio of coopera-tion, collaboration, conflict, and negotiation in which morally laden interaction logics determine the more socially rewarding courses of action.

The triadic unit of analysis has two domains of interaction: the domain of within- interactions that accounts for the interactions that hold together humans and artifacts inside collectives and the domain of between- interactions that ac-counts for the interactions between collectives (e.g., cooperation, collaboration, or conflict).1 Both domains are intrinsically related because within- interactions assemble meanings and mediate intentionalities reified by between- interactions. For simplicity, in the remaining text I use the bracketed notation [a- b] for within- interactions and a←→b for between- interactions.

A dissection of within- interactions requires a closer look at two facets of people interacting with artifacts. On facet accounts for how meaning is assembled; the other considers how artifacts mediate human intentionality. Latour (1999, 178– 85) characterizes the assembly of meaning by describing how artifacts and humans translate, compose, pack, or delegate their programs- of- action.2 His proposal has the advantage of focusing on the mechanisms of meaning articulation rather than on the meaning itself. The other facet accounts for how artifacts mediate human intentionality. It comprehends both the mental state of us and the object of such mental state. Husserl terms the former “noesis” and the latter “noema” (Dourish 2001). For Ihde (1990), people’s mediated interactions with the world always occur through technologies that are present to someone in different degrees ac-cording to where it is located within our object of intentionality (Ihde 1990; Ver-beek 2005, 2015). Such degrees of presence describe a continuum: On one side, the artifact’s mediation of human experience of the world is unnoticeable (e.g., the car

[human – artifact] other

Within-interactionAssembly of meaningMediation of intentionality

Between-interactionCooperationCollaboration

FIGURE 1: Triadic unit of analysis and domains of interaction. The within- interaction domain exhibits two facets: assembly of meaning and mediation of intentionality. The between- interaction domain com-prises the forms of social interaction between collectives.

502 SALAMANCA

suspension operates in the background while one is driving). On the other side, the artifact’s mediation is fully present in the foreground of human awareness (e.g., the speed bump hindering the car- driver flow). In the domain of within- interactions when the mediating artifact is unnoticeable, the object of intentional-ity targets actors outside the human- artifact collective. Conversely, when the mediating artifact is present- at- hand, the object of intentionality targets actors in-side such a collective.

To illustrate how within- interactions can be dissected in relation to their two facets, picture a driver at an intersection switching on the car’s left turn signal to warn a pedestrian of her intention. On the one hand, two forms of the assembled meaning are evident in that situation: composition because the unit [car- driver] turning to the left is not just the driver but a composed meaningful collective, and delegation at the moment the driver delegates to the car the function of warning the pedestrian of her intention. On the other hand, the mediated intentionality is explicit as the car embodies a unit with the driver, and this unit’s object of inten-tionality targets the pedestrian, who is outside the domain of the collective’s within- interaction.

Empirical Case Study

In order to test the proposed methodological instrument, I designed an empirical comparative study of pedestrians walking on a street crosswalk. In this section I explain the procedure and findings of an examination of raw within- interactions and between- interactions in the wild that sets the touchstone by which a custom- designed smart crosswalk was further evaluated. In the following section I de-scribe such evaluation in terms of how the new smart crosswalk affected cooperation and collaboration in staged interactions of pedestrians. As mentioned earlier, the smart crosswalk is not intended to be a product available on streets; it is rather a research tool for the exploration of how smart artifacts affect emergent social interactions.

The case chosen for study were the naturally occurring interactions on a busy downtown crosswalk in the city of Chicago. Crosswalks are illustrative venues of nonverbal emergent cooperation, collaboration, conflict, and negotiation, where cliques and single walkers exhibit processes of participatory sense making. The sources of evidence were direct observation and two hours of video clips that cap-tured dozens of walkers interacting on the crosswalk from a bird’s- eye point of view. The video footages were analyzed with Path Analytics, a custom- made video analysis tool that generated a detailed dataset of each pedestrian path including time- stamped strides, course, displacement, and deviation from expected trajec-tory (see figure 2). It also visualizes forecasted positions and potential zones of conflict between pedestrians and cliques.

The analysis showed that urban pedestrians walk by themselves or in cliques of two or three. Larger groups of acquaintances regularly split in combinations of pairs or trios. The configuration of walking dynamics on sidewalks can be sum-marized in six combinations of walker sets (single, pair, or trio) with opposing trajectories— named “north pedestrians” and “south pedestrians.” The analysis also showed that the crosswalk width and the size of the interacting cliques have a strong effect on the overall walking flow. Large cliques exhibit steadier courses, whereas small ones often divert their trajectories.

503SMART ARTIFACTS

The moment pedestrians step on the crosswalk they subscribe to the program- of- action inscribed in its design attributes (i.e., afford walkers traveling straight across corners) constituting a collective. The representative triadic unit of analy-sis is defined as the collective of north pedestrians walking of the crosswalk in-teracting with south pedestrians or [north pedestrians– crosswalk] ←→south pedestrians. This unit is simple enough to illustrate the complexity of the triadic interaction, and rich enough to depict the human and nonhuman interactions of walkers on the street. In terms of the two facets of the within- interaction domain, the facet of assembled meaning is a composition— in Latour’s terms— resulting in the articulation of pedestrians and crosswalk programs- of- action. As for its medi-ated intentionality, the locus of each walker’s intentionality is not at the approach-ing pedestrian nor at the crosswalk but at the foreseen conflicting intersection of the three programs- of- action reified in a location on the pattern of white stripes. The analysis of the between- interaction clearly revealed that people walking in the same direction collaborate whereas those heading in the opposite direction coop-erate. This triad of actors dissolves at the moment the pedestrians step out of the crosswalk unsubscribing from its program- of- action. Collaborative interactions differ from cooperative ones in that in the former two or more actors or collectives align their programs- of- action to achieve congruent goals, whereas in the latter they weave their discordant programs- of- action while pursuing their own best possible outcomes. Collaborative parties have selfless interest in contributing to the achievement of the goals of other actors. This was observed when cliques aligned themselves to follow the trail of unacquainted predecessors. Coopera-tive parties may have selfish attitudes and share a symbiotic interaction be-cause their mutual successes benefit one another. This was frequently observed when single walkers with conflictive trajectories with large cliques gave them the right of way.

FIGURE 2: Analysis of trajectory interaction between pedestrians. The clear areas describe the zone of cooperation of a clique of three walkers that open space for another one walking in opposite direction. The hatched areas are the potential zones of conflict between them.

504 SALAMANCA

Staged Studies and Smart Crosswalk Design

Due to the fact that walkers locate their intentionality at the points of intersection foreseen by themselves in the white striped pattern, I designed an intelligent pe-destrian crossing that predicts such common points of intersection and visually points out ways to avoid the potential conflict. For the prediction, a computer video processor anticipates in real time spots of conflicting interaction from a bird’s- eye video source located above walkers on the smart crosswalk. Those spots became the splitting epicenters of the white striped pattern that timely prompted changes in walking courses by signaling the most balanced distribution of walking space among concurrent cliques (see figure 3). The signaling behavior had two modes: In the subjunctive mode the smart crosswalk signaling tone is a suggestion of the ideal distribution of its width proportionally to the size of the cliques.3 It simply opens a gap in the striped pattern along the crosswalk longitudinal axis. In the prescriptive mode, the signaling tone is indicative, also displaying a balanced distribution of the walking space but emphasizing a walking direction. In this mode, the smart crosswalk splits the stripped pattern along a diagonal axis and slightly slides both halves in opposite direction indicating both parties a trajectory to get around the zone of conflict. In some of the trials the walkers’ visual perception range was lim-ited, expecting that they would heavily rely on the information signaled by the smart crosswalk’s pattern.

In a staged study, 48 walkers organized in cliques of one, two, and three ran 12 trials to test a full- sized smart crosswalk deployed at the IIT Institute of Design in Chicago (see figure 4). With the help of Path Analytics, I recorded the same vari-ables as those observed in the wild and contrasted them.

South end

No-signal

Subjunctive

Prescriptive

North end

No-signal

Subjunctive

Prescriptive

FIGURE 3: Smart crosswalk signaling modes tested in a staged experiment.

505SMART ARTIFACTS

Results and Interpretation

For the most part, both smart crosswalk signaling modes positively impacted the consistency of pedestrians’ walking course because their trajectory deviation was considerably lessened in comparison with observations in the wild. This con-stitutes evidence of improvement in coordination. Yet, in spite of better coordina-tion, the pedestrians’ walking speed was slower. A possible explanation is that pedestrians do not care about how much they need to deviate from their course as long as they maintain their pace. Hence, I deduce that cliques prioritize the achievement of their intended target at a consistent velocity compromising their course stability.

The subjunctive signal produced lower target deviations and clearer pedestrian walking paths than the prescriptive signal. An interpretation is that the pedes-trians might improve the coordination of their actions when their attention is directed toward potential conflicts. In other words, smart crosswalk signals com-plement pedestrians’ contextual information of the ongoing social situation. How-ever, pedestrians’ efficiency to circumvent those conflicts was diminished when prescriptive signals were suggested.

The limitation of walkers’ visual perception yielded no surprising results. Pedes-trians walk faster in full visual perception, and the shorter the visual perception field, the slower the walking velocity. Similarly, the larger the number of subjects participating in a trial, the slower their average speed. But data show intriguing results when we scrutinize the walkers’ trajectory deviations because people with full visual perception, both in the staged experiment and on the street, tend to di-vert their trajectories significantly more than those with limited visual percep-tion. A careful analysis of clique members’ walking course revealed that one

FIGURE 4: Staged experiment of smart crosswalk signaling patterns.

506 SALAMANCA

member usually makes a greater effort to maintain her trajectory flow. In full vi-sual perception, the larger clique yields the right of way over the opposing party when the smart crosswalk is in subjunctive tone, whereas the opposite was ob-served in prescriptive tone. An interpretation is that calling the attention of pedes-trians to their potential conflicts via the subjunctive signals results in improvements of the walking flow of the smaller group. Consequently, the largest party has to exert greater effort to maintain its course. The inverse situation is observed in pre-scriptive signal. The larger collectives benefit the most from the smart crosswalk prescriptions, shifting the burden of adapting the trajectory to the smaller collectives.

An examination of the within- and between- interactions of the triad [pedestrian- smart crosswalk] ←→ other pedestrian helps us to elucidate the creative pro-cess of augmenting a traditional artifact with autonomous and adaptive functionalities that enable it to actively intervene in social practice. Let us make first a parallel between the within- interaction domains of regular crosswalks and smart crosswalks. In either a regular crosswalk or a smart crosswalk, the assembled meaning is the result of a composition of actors, but the smart crosswalk packs more actors than the regular crosswalk because it not only includes tangible components but also entails a designer/programmer of nonhuman attitudes evident in the sig-naling tone of the smart crosswalk. In terms of walkers’ intentionality mediation, there is a radical shift from an unnoticeable mediation of the regular crosswalk to a hermeneutic mediation of the smart crosswalk that combines visual and perform-ing languages to overtly signify frequent courses of action for the common good. Simply put, a smart artifact is a signifier of morally accepted situated actions.

As for the between- interaction domains of regular crosswalks and smart cross-walks, cooperation and collaboration emerged indistinctively, but the coordination among walkers and the stability of their walking paths are remarkably improved while walking on smart crosswalks, especially in prescriptive mode. This means that smart crosswalks benefit the cohesion and stability of large cliques.

The Social Viscosity of Assemblages

The use of the triadic unit of analysis in the staged design study reveals that while actors collectively attempt to enact their programs- of- action, they offer resistance to one another, hindering the interaction flow of the entire assemblage. Following the naïve definition of mechanical viscosity— the resistance to flow exhibited by a viscous matter— I define social viscosity as the resistance to social action flow elicited by human and nonhuman actors acting concurrently.

Actor- networks exhibit social viscosity that affects the flow of independent ac-tion unevenly. While smaller collectives coordinate easily, larger collectives strug-gle to establish and maintain coordination by continuous acts of balancing and rebalancing. The observed accumulated disturbance reveals the internal friction of actors, as a result of their attempt to enact their programs- of- action. Such fric-tion, which ultimately renders the actor- network viscous, seems to thicken when people act under limited access to prescient environmental information. It is under those limited conditions when smart artifacts may have a greater impact on the network’s viscosity and benefit communal action flow. With the example of a skillful doorman and a mechanical door closer, Latour (1988) exemplifies how highly trained humans and better designed nonhumans could facilitate the flow of action

507SMART ARTIFACTS

in an actor- network. In the same vein, this study demonstrates that the design of socially apt nonhumans associated with well- behaved humans could reduce the social viscosity of the actor- network.

Conclusion

The embrace of autonomous and adaptive everyday artifacts by society highlights the social dimension of interaction design. Instead of regarding everyday artifacts as mere tools or equipment, they need to be considered as agents with capacity for action in social contexts.

This research adopted a posthumanist position to tackle such conceptual shifts. It repositions artifacts within social practices as active participants symmetrical to humans enrolled in such practices. The triadic structure of networked social inter-action presented herein is a methodological instrument for the investigation of smart devices adaptively mediating people’s interactions. Such a triadic unit of analysis ac-counts for the interactions within and between collectives in actor- networks. The within- interactions are those that hold together humans and smart artifacts inside a collective and put forward the collective’s meaning and intentionality for other actors in the network. The between- interactions are those that occur among collectives and characterize the dominant model of sociality of the actor- network.

I envision that the triadic structure can be used in the interaction design pro-cess as an instrument for analysis of activities and identification of their focuses of activity. Moreover, the proposed triadic structure is an instrument for the eval-uation of the social effects of artifacts in an actor- network in that the study of the between- interactions reveals the types of social interaction in the triad— cooperation and collaboration in this case— describing the directionality of actors’ goals, alignment of their programs- of- action, and mutual consideration for each other’s interests. The analysis of cooperation and collaboration demonstrated that coordination is a contingent condition for their emergence, regardless of the pres-ence of coordination technologies. But our findings indicated that any mediator, at the expense of the fluidity of social action, may positively affect the degree of coor-dination. The challenge for smart artifact designers is to design social mediators that enhance— and do not hamper— the fluidity of social action.

This research suggests that researchers and practitioners designing smart artifacts aiming to promote cooperation and/or collaboration need to do the following:

1. Define the social practice in which the smart artifact will be involved and select the most relevant activities.

2. Understand the set of social logics and values that define the social domain of the practice and the cultural behaviors of human participants that engage and affect the practice.

3. Map the triads of networked humans and nonhumans and describe their constituent collectives. Identify which artifacts have the highest potential to become smart artifacts by determining which nonhuman actors are located at relevant focuses of activity.

4. Analyze the within- interactions of collectives to understand the assembled meanings by technological means and the mediation of intentionalities of their hybrid actors.

508 SALAMANCA

5. Analyze the between- interactions to understand the aspects that define the interactions between collectives: directionality of goals, compatibility of programs- of- action, and consideration for others’ interests. In doing so, spec-ify the type(s) of coordinated action as cooperative or collaborative.

6. Design smart artifacts that signify the social logics and values, and promote types of coordination that support between- interactions.

7. Deploy smart artifacts in the social practice and evaluate their impact in the social viscosity of the actor- network.

Notes

1. A collective is defined as an association of heterogeneous actors that forms a meaningful whole within an array of human activity (Latour 1999, 2005).

2. Translation is exemplified by how the moral integrity of a good citizen is degraded when she holds a gun; composition is exemplified by how traveling by plane is possible thanks to the association of pilots, planes, airports, launch pads, air traffic controllers, and ticket counters; black- boxing is ex-emplified by the network of repairmen, light bulbs, and lenses packed in an overhead projector (i.e., a sealed unit) that gets unpacked when the device breaks down in the middle of a lecture; and finally, delegation is exemplified by traffic engineers entrusting the reduction of car drivers’ speed to speed bumps.

3. On subjunctive and prescriptive interaction, see Tomasello (2010) and Sennett (2012).

Works Cited

Antonelli, Paola. 2011. Talk to Me: Design and the Communication between People and Objects. New York: Museum of Modern Art.

Callon, Michel, and John Law. 1995. “Agency and the Hybrid Collectif.” South Atlantic Quarterly 94 (2): 481– 507.

DiSalvo, Carl. 2012. Adversarial Design: Design Thinking, Design Theory. Cambridge, MA: MIT Press.Dourish, Paul. 2001. Where the Action Is: The Foundations of Embodied Interaction. Cambridge, MA: MIT

Press.Dunne, Anthony. 2005. Hertzian Tales: Electronic Products, Aesthetic Experience, and Critical Design. Cam-

bridge, MA: MIT Press.Dunne, Anthony, and Fiona Raby. 2013. Speculative Everything: Design, Fiction, and Social Dreaming. Cam-

bridge, MA: MIT Press.Forlano, Laura. 2016. “Decentering the Human in the Design of Collaborative Cities.” Design Issues 32 (3):

42– 54.Hutchins, Edwin. 1995. Cognition in the Wild. Cambridge, MA: MIT Press.Ihde, Don. 1990. Technology and the Lifeworld: From Garden to Earth. Indiana Series in the Philosophy of

Technology. Bloomington: Indiana University Press.Knorr Cetina, Karin. 1997. “Sociality with Objects.” Theory, Culture & Society 14:1– 30. doi:10.1177 /

026327697014004001.Kuutti, Kari, and Liam J. Bannon. 2014. “The Turn to Practice in HCI: Towards a Research Agenda.” In

Proceedings of the 32nd Annual ACM Conference on Human Factors in Computing Systems, 3543– 52. New York: ACM.

Latour, Bruno. 1988. “Mixing Humans and Nonhumans Together: The Sociology of a Door- Closer.” So-cial Problems 35 (3): 298– 310.

———. [1991] 1993. We Have Never Been Modern. Edited by Nous n’avons jamais ètè modernes. Cambridge, MA: Harvard University Press.

———. 1999. Pandora’s Hope: Essays on the Reality of Science Studies. Cambridge, MA: Harvard University Press.

———. 2005. Reassembling the Social: An Introduction to Actor- Network- Theory. Oxford: Oxford University Press.

509SMART ARTIFACTS

Maturana, Humberto. 2002. “Autopoiesis, Structural Coupling and Cognition.” Cybernetics & Human Knowing 9 (3– 4): 5– 34.

Norman, Donald A. 2002. The Design of Everyday Things. New York: Doubleday.———. 2011. Living with Complexity. Cambridge, MA: MIT Press.Norman, Donald A., and Stephen W. Draper. 1986. User Centered System Design: New Perspectives on Human-

Computer Interaction. Hillsdale, NJ: Erlbaum.Porter, Michael E., and James E. Heppelmann. 2014. “How Smart, Connected Products Are Transform-

ing Competition.” Harvard Business Review 92 (11): 64– 88.Ratto, Matt. 2011. “Critical Making: Conceptual and Material Studies in Technology and Social Life.”

Information Society 27 (4): 252– 60.Rosenberg, Karen. 2011. “Art That Interacts if You Interface.” New York Times, July 29. www.nytimes

.com/2011/07/29/arts/design/momas-talk-to-me-focuses-on-interface-review.html?_r=0.Salamanca, Juan. 2012. “Designing Smart Artifacts for Adaptive Mediation of Social Viscosity: Triadic

Actor- Network Enactments as a Basis for Interaction Design.” Doctoral dissertation, Illinois Insti-tute of Technology. https://pqdtopen.proquest.com/pubnum/3570096.html.

Schatzki, Theodore. 2001. “Practice Theory.” In The Practice Turn in Contemporary Theory, edited by Theodore Schatzki, Karin Knorr Cetina, and Eike Savigny, 1– 14. Oxon: Routledge.

Sennett, Richard. 2012. Together: The Rituals, Pleasures, and Politics of Cooperation. New Haven, CT: Yale University Press.

Serrano, Manuel M. [1977] 2008. La Mediación Social. Madrid: Akal.Suchman, Lucille Alice. 2007. Human- Machine Reconfigurations: Plans and Situated Actions. 2nd ed. Cam-

bridge: Cambridge University Press.Tomasello, M. 2010. ¿Por qué Cooperamos? Buenos Aires: Katz Editores.Verbeek, P. 2005. What Things Do: Philosophical Reflections on Technology, Agency, and Design. University

Park: Pennsylvania State University Press.———. 2015. “Cover Story: Beyond Interaction: A Short Introduction to Mediation Theory.” Interactions

22 (3): 26– 31. http://dx.doi.org/10.1145/2751314.Vinck, Dominique. 2012. “Accessing Material Culture by Following Intermediary Objects.” In An Ethno-

graphy of Global Landscapes and Corridors, edited by Loshini Naidoo, 89– 108. London: Intech.